How you set up your network for voice, data, video, and audio transmission can mean the difference between a network that operates efficiently and transparently to users and one that’s slow and tedious.

Think of your network as a river of data. You have a steady current of data moving smoothly down the channel. All your network users are like tiny tributaries branching off the main river taking only as much water (bandwidth) as they need to process data. When you start to multicast voice, video, data, and audio over the LAN, those streams suddenly become the size of the main river. The result is that each user is basically flooded with data and it becomes almost if not impossible to do any other tasks. This scenario of sending transmission to every user on the network is called broadcasting, and it slows the network down to a trickle. But there are network protocol methods that alleviate this problem.

Unicasting vs. multicasting

Unicasting is sending data point-to-point, from one network device to another. Multicasting is transmitting data from one network device to multiple users.

When multicasting with Layer 2 switches, all attached network devices receive the network packets, whether they want them or not. When you multicast with Layer 3 switches (with multicast support), you send the network packets to only the exact client/receives who want them. You control where the river of data goes and put up locks to keep the river out of other user streams.

Unicasting technical details

Every device on a network has a MAC address and an IP address. In a typical network, we usually see a lot of unicast packets going from one device to another. The unicast packet header will contain the destination IP address of the device it needs to reach and typically resembles a Class C IP address ranging from 192.0.0.0 all the way to 223.255.255.255. These unicast packets are generated from a single source and are sent to a single receiving device. The source nodes, also known as network switches and routers, have a built-in routing table, which enables the unicast packet to find its destination by taking the shortest hop from one switch to another.

In a typical unicast network, you’ll find that Layer 2 switches are used because they’re very cost effective and can easily support these types of network communications.

Multicasting: Layer 2 vs. Layer 3 switches

In multicasting, a single device will transmit a network packet out to many, but not all, devices that are actively tuned in and want to receive the packet. A multicast IP address is a Class D address, which means it has a range of 224.0.0.0 to 239.255.255.255. The IP address range typically designated for multicasting protocols is between 224.0.0.0 through 224.0.0.255.

Multicasting with Layer 2 switches. Because a multicast head does NOT have a destination IP address, an average network switch (a Layer 2 switch without supported capabilities) will not know what to do with it. So the switch sends the packet out to every network port on all attached devices. When the client or network interface card (NIC) receives the packet, it analyzes it and discards it if not wanted.

Multicasting with Layer 3 switches. To solve the problem of sending packets to every device on the network, use a Layer 3 switch that supports IGMPv2 or IGMPv3 and packet forwarding (Internet Group Management Protocol [IGMP] supports multicasting on the physical network level).

The Layer 3 switch multicasts network packets only to devices that have tuned into the stream. It does not send the multicast packets to devices that don’t want it. This makes multicasting with Layer 3 switches much more efficient than with Layer 2 switches.

The value of IGMP and Layer 3 switches

If you had a Layer 2 switch that didn’t support IGMPv2 or IGMPv3 snooping, the switch would be able to handle only a few devices sending multicasting packets.

Layer 3 switches with IGMP support, however, are smart enough to know who wants to receive the multicast packet and who doesn’t. When a client/receiving device want to tap into a multicasting stream, it responds to the multicast broadcast with an IGMP report. That’s the equivalent of saying, “I want to connect to this stream.” The IGMP report is only sent in the first cycle and initializes the connection between the multicast stream and receiving device.

If the client/receiving device was previously connected to the stream, it sends a grafting request to the Layer 3 switch. Grafting enables the IGMP-supported switch to remove the temporary block on the unicast routing table. The switch can then send the multicast packets to newly connected members of the multicast group.

When the client/receiving device no longer wants to receive the multicast packets, it sends a pruning request to the IGMP-supported switch, which temporarily removes the device from the multicast group and stream.

Basically what we’re saying is…

If you plan on using multicasting protocols on your network, make sure you have routers or Layer 3 switches that support the IGMP protocol. Without this support, your network devices will be receiving so many multicasting packets, they will not be able to co communicate with other devices using different protocols, such as FTP.

Also, if you’re planning to send video from multiple sources over a LAN, a feature-rich, IGMP-supported Layer 3 switch also will give you the bandwidth control you need.

Be sure to configure the IGMP-supported switches for the particular application. Most Layer 3 switches right out of the box have IGMP support disabled.

Resources

Global Centralized KVM Server Management White Paper

Video: Go KVMoIP to save time, money

KVM Selector (A tool to configure your ultimate ServSwitch KVM solution)

Is your company concerned about being connected to the Internet? If  so, you’re not alone. It’s the fear of being successfully hacked that’s  keeping many companies and their employees off the Internet. In fact,  according to Infosecurity News, over 25% of all companies polled  responded that hackers have tried to gain unauthorized access from  outside the corporate network.

What else keeps companies disconnected? Hackers are no longer just  techies with a hobby. Hackers are no often highly trained professionals  in places like Eastern Europe, Russia, or China, and they’re out to make  a profit on you. There’s a mystique to being hacked. It’s the new crime, it’s the hottest news. And it’s fun…according to hacking group Lulz Security, whose motto claims they are, “The world’s leaders in high-quality entertainment at your expense.” Well  they sure aren’t fun for you. If anything, this hype from the media  watchers and technical experts, should be keeping you on your toes.

So why the mystique? Our ever-growing, ever-faster cyberworld makes  protecting your network from intruders ever more difficult. Many people,  including company executives,  lack a fundamental understanding of technology. Mix this with the  complexities of network security issues and you can see why many  companies haven’t ventured onto the Information Superhighway. However,  what you don’t know really can hurt you.

On the other side of the spectrum, there are many companies with  executives who have “Teflon sensibilities.” Media hype doesn’t stick to  them. Unfortunately, neither does the advice of company network  administrators who want network security problems addressed. These  companies go about their connected-to-the-Net business until the  inevitable compromise in security happens. And when something valuable  is swiped, the executives worry.

“We’ve taken steps to make sure something like this never happens  again!” is the boilerplate reaction. Otherwise, they say, heads will  roll. For companies that have been hacked, the cost of information security now equals the cost of the incident plus the cost of countermeasures.

So protect yourself from hackers while getting the Internet access  you need. You’ll never be 100% secure, but you can dramatically reduce  your risk and proactively defend your organization by containing and  controlling threats, vulnerabilities, and assets. Just use the 4Ds:

• Threats need to be detected, deterred, defended against, and defeated in real-time or expect downtime.
• Vulnerabilities need to be detected, deterred, defended against, and defeated (i.e. removed by system hardening, reconfiguration, patching, etc.) as quickly as possible or expect to be exploited.
• Assets need to be controlled—which ones gain access to your  network/infrastructure and those that are trusted but weak or infected  need to be quarantined in real-time or expect malware propagation.

It’s a fine balancing act, but you can protect data and keep  out the unwanted while still giving your staff access they need. A  network, as it turns out, is only as secure as the people who run it,  use it, and fund its protection. Find out more about protecting your  data infrastructure at www.blackbox.com/go/Security.